What does the CAP theorem imply in a partitioned distributed ERP system?

• A. Consistency models are independent of partition.
• B. Availability must be sacrificed to maintain consistency.
• C. Consistency must be sacrificed to achieve availability.
• D. In the presence of a partition, you must choose between consistency and availability.

Answer: (D)

CAP theorem shows that in a distributed ERP system dealing with a network partition, you can’t have both consistency and availability at the same time. When the network splits, you must pick one: either maintain strict consistency, which may require delaying or rejecting requests so all nodes agree on the latest state, or maintain availability, which lets the system respond immediately but can return data that isn’t fully synchronized across nodes.

In practice, this means if you prioritize consistency (CP approach), you ensure all reads reflect the same latest write, which is critical for financials and accounting, but some operations may stall or fail during a partition. If you prioritize availability (AP approach), the system keeps responding to users even during a partition, which keeps processes like order entry moving, but data can become temporarily divergent and requires reconciliation after the partition heals.

The key takeaway is that, when a partition exists, you must trade off between keeping data perfectly consistent and keeping the system responsive.